In the pig industry, male piglets are surgically castrated early in life to prevent boar taint. Boar taint is mainly caused by androstenone and skatole. Androstenone is a pheromone that can be released from the salivary glands when the boar is sexually aroused. Boars are housed in groups and as a consequence boars can influence and be influenced by the phenotype of other boars by (non-)heritable social interactions. The influence of these social interactions on androstenone is not well understood. The objective of this study is to investigate whether androstenone concentrations are affected by (non-)heritable social interactions and estimate their genetic correlation with growth rate and backfat. The dataset contained 6,245 boars, of which 4,455 had androstenone observations (68%). The average number of animals per pen was 7 and boars were housed in 899 unique pen-groups (boars within a single pen) and 344 unique compartment-groups (boars within a unique ‘room’ within a barn during time). Four models including different random effects, were compared for androstenone. Direct genetic, associative (also known as social genetic or indirect genetic effects), group, compartment, common environment and residual effects were included as random effects in the full model (M3). Including random pen and compartment effects (non-heritable social effects) significantly improved the model (M2) compared with including only direct, common environment and residual as random effects (M1, P <0.001), and including associative effects even more (M3, P <0.001). The sum of the direct and associative variance components determines the total genetic variance of the trait. The associative effect explained 11.7% of the total genetic variance. Backfat thickness was analysed using M2 and growth using M3. The genetic correlation between backfat (direct genetic variance) and total genetic variance for androstenone was close to 0. Backfat and the direct and associative effects for androstenone had genetic correlations of 0.14 ± 0.08 and -0.25 ± 0.18, respectively. The genetic correlation between total genetic variances for growth rate and androstenone was 0.33 ± 0.18. The genetic correlation between direct effects was 0.11 ± 0.09 and between associative effects was 0.42 ± 0.31. The genetic correlations and current selection towards lower backfat and greater growth rate suggest that no major change in androstenone is expected when breeding goals are not changed. For selection against boar taint and therefore also against androstenone , results recommend that at least the social environment of the boars should be considered.